Chess pieces

AI doesn’t change your strategy – it changes how you deliver it

Strategy Director Emma Andrews lifts the lid on how the introduction of AI has impacted Luminary's own strategy.

Emma Andrews

18 May 2026

4 minute read

The word ‘strategy’ is often synonymous with big, new ideas, innovation or just generally something fun to do that people want to be a part of. It’s put on a pedestal for that reason.

Similarly in Roger Martin’s Strategy Choice Cascade, which is the framework we use at Luminary, it’s easy for most of the attention to go to the first three choices:

  • Winning aspiration
  • Where to play, and
  • How to win.
Roger Martin's Strategy Cascade

They’re the interesting parts. The parts that spark debate. The parts that feel like strategy.

Where organisations often lose energy is further down the cascade – in the choices that actually make strategy real:

  • What capabilities must be in place to win?
  • What management systems are required to build and sustain those capabilities?

That’s the work of execution. And it’s where strategy either holds – or quietly unravels.

What’s changed (and what hasn’t)

At Luminary, AI hasn’t fundamentally changed our winning aspiration, where we play, or how we aim to win.

What it has done is force us to take a much harder look at the capabilities and management systems required to deliver on that strategy. Because AI doesn’t sit alongside an organisation like Luminary; it interacts within it, determining:

  • how work is structured
  • how processes are delivered, and
  • how outcomes are created.

And that means it places pressure on the systems that already exist.

Rethinking capability: from broad to deliberate

One of the first things we’ve had to confront is the way we define capability. It’s easy – and common – to create long lists of things an organisation needs to be good at. And that long list probably wouldn’t look that different to the next agency on the client’s shortlist.

In practice, not all capabilities are equal. The ones that matter are the few that help you win where you have chosen to play. That is, they:

  • underpin your ability to deliver value
  • are invested in disproportionately, and
  • are performed distinctively.

With AI, that distinction becomes sharper.

Rather than asking: Where can we use AI? We’ve been asking:Which of our core capabilities does AI strengthen – and where does it require us to revisit our traditional processes?

If considered as a precious resource rather than a silver bullet, this means we’ll be more deliberate and discerning, clarifying where human judgement remains critical, identifying where AI can provide speed, consistency or insight, and defining how those two work together as a single capability, not separate ones.

Rethinking management systems

If capability is what you need to do well, management systems are how that happens consistently. Before introducing AI more broadly, much of this sits implicitly in organisations:

  • workflows evolve over time
  • decision-making is often experience-based, and
  • information is tacitly held.

AI doesn’t remove that variation and complexity – it exposes it.

So rather than layering AI on top, we’ve been forced to make those systems more explicit. That has meant defining clearer workflows before introducing automation. Once we’d improved how information was structured, captured and accessed then we could add the AI layer alongside clear governance principles for how it should and should not be used. In effect, AI has required us to formalise parts of the system that previously relied on tacit knowledge.

AI as a capability rather than a tool 

The biggest shift we need to make is from treating AI not as a set of tools, but as an organisational capability in its own right. This means developing internal frameworks to guide where and how AI is applied and only embedding AI into workflows when it supports, rather than replaces, human decision-making. To achieve this we’re piloting the priority use cases and embedding guardrails to ensure consistency, quality and appropriate use. This is less about adoption, and more about integration. Because without that structure, AI risks sitting alongside the organisation rather than strengthening it.

Where this leaves strategy

AI hasn’t changed what we are trying to achieve. But it has changed what it takes to get there. And in doing so, it has shifted our attention to the parts of the strategy cascade that are easiest to overlook – but most critical to get right. Given the pace of change, we expect to revisit them again in six months as we head into 2027. Because that’s where strategy becomes something you can actually deliver.

Download our Digital Strategy Roadmapping Toolkit

Download now

Keep Reading

Want more? Here are some other blog posts you might be interested in.